This is 2035 — not 2050. We are probably already at peak oil and the decline of fossil fuels will decrease nonlinearly due to compounding improvements in solar and batteries. The extent that some countries can escape this through diversification (UAE) is a relevant and interesting question.
Worth pointing out that in many cases, a country’s lack of rare earth mineral refining comes from deliberate environmental commitments. The US did used to have its own mines and refineries; we shut them down due to pollution and climate commitments. But the minerals are still in the ground. You can bet as lack of REEs becomes more and more of a national strategic threat, Western governments will adjust their standards.
Excellent article - @Tomas Pueyo is among the limited (but growing) number of commentators who really get that the end of the fossil fuel era is now economically inevitable, and this is a valuable geopolitical projection extrapolating from that.
It's becoming increasingly obvious how dependance on oil fuels conflict.
With respect to rare minerals it is the processing that China has taken control of in part by doing it at a loss without regard for their own environment so almost no one can compete. Countries like Canda that have the minerals should be investing in finding ways to process it without the environmental disaster it is right now.
From idea to functioning plant(s) and mine(s) is definitely not a 2 year thing, 5 years maybe but that's if we drop everything and get onto that. China can do the 2 years but they can do things at a loss without concern for the environment or what people want or don't want built near them.
I really value your opinion, as I've mentioned multiple times, but, in this case, I think you're the one doing a lot of wishful thinking, especially in your 2050 geopolitical predictions. And I'm really failing to understand your prediction for Angola. First of all, ethnic conflict in Angola is practically non-existent since the end of the civil war. Also, according to v-dem and eiu, it's currently an hybrid regime, not an authoritarian state (like your graph says). Second, its internal skirmishes are miniscule when compared to Nigeria and Chad (both in number of casualties and the fact that they don't have the Islamic State on their doorstep). Also, the conflict is exclusive to the exclave of Cabinda, which happens to have all the oil reserves, and that's the big reason there's a fight. If the value of oil is gone, Cabinda has even less reason for fighting.
Angola has a lot of problems and, if the value of oil really falls like you say, Luanda might lose its status as one of the richest and most developped cities in sub-saharan Africa, but I don't see the country plummeting into a Civil War (again); and also don't see them becoming a "strong democracy", out of nowhere. If things go well, they'll simply continued slowly democratizing, like they've had in the last 20 years.
I am indeed quite hopeful, and my optimism biases my hopes. Not sure what forecast is wishful thinking though? The intro is not a prediction, it’s a potential scenario. And I actually don’t wish that! Too much war in there.
On Angola, it’s a country I haven’t dived deeply into yet so I might get it wrong. But its exposure to oil is objectively high, given the cost of its extraction and the share of gov income it represents. Do you really think the country would remain unscathed with a drop in oil demand?
More importantly, I am not arguing that Angola is very exposed mainly because of its existing internal strife. But by the oil exposure.
I can't imagine Iran losing territory to Azerbaijan. I can't see Turkey allowing the emergence of an independent Kurdistan, and much less one where the Iranian Kurds gained influence over the other kurds of the region. The Iraqi Kurds are the only ones with any sort of autonomy and they seem really keen on peace.
Also, as someone who's got to meet a few immigrants in Saudi Arabia, I think you overestimate the fragility of its governance, so I also don't see Hejaz seceding from it (even with the arguments given in the very insightful article on it).
Syria being at peace with Israel seems extremely hard to achieve (much less in a defense pact under them), because Israel is never gonna trust them enough to give them the Golans back and they simply won't accept that. I don't know how their story end, but I'm pretty sure it's not with Syria bowing to Israel.
And why would Dubai, with the strongest passport in the world, one of (if not the) most important airport, and one of the highest soft power capabilies in the world, throw away its stability and risk giving it all up for a petty war? Even if they had anything to gain by it, the losses they would get from the war would make them reconsider.
As for Angola, a drop in oil demand will definitely affect the country, but I think only Luanda and Cabinda would notice. And I simply don't see how that could ever lead to a civil war. Who would be fighting whom? For control of what?
Yes, of course none of that is a prediction! This is one of millions of potential scenarios! The idea is to paint a picture of the mess that could happen with the end of oil, to make it real. And I sure believe that's the case: The transition will be messy!
Syria is at peace with Israel at this very moment.
How many people predicted a (bloody, long) civil war in Libya in 2010? Heck, how many people predicted a (bloody, long) civil war in Syria in 2010? Or Yemen?
The break up of a bunch of countries in the Middle East actually isn't as unlikely as you seem to think (granted, the various emirates of the UAE going to war with each other seems a bit farfetched).
I’m a mega supporter of nuclear, but it takes a long time to be built. Regardless, whether it’s through solar or nuclear, the reaction to oil doesn’t change.
There are several tax haven countries that provide the equivalent of pumping regulatory arbitrage out of the ground. The smaller petro states could become this, but the emphasis is on small. You can't share the resource with tens of millions of people.
I think plastics are something like 6% of oil and none of gas? If so, dividing by 20x the amount of oil sales in the world is close enough to eliminating it in geopolitics terms.
the diversification framing undersells what's actually happening. when MGX targets $100bn in AI infrastructure and Saudi PIF combines $240bn for data centers, the story isn't petrostates escaping rent-seeking, but petrostates porting the rent model to a new layer.
oil worked because it was physically concentrated, capital-intensive, and infrastructure everybody had to use. compute is shaping up identically. the gulf states know the playbook because they built the last version of it.
norway escaped the resource curse by investing the rent into diversified global assets. most of the others are shopping for the next extractable moat. i wrote about the AI landlord structure (https://huliwood.substack.com/p/the-ai-landlords). your 2050 scenario adds the layer i'd underweighted: the most dangerous outcome isn't petrostates failing to diversify. rather: petrostates discovering that compute IS the new oil, and they already have the capital to buy the wells.
There is no way AI can be exploited in the same way as oil, because datacenters can move (and depreciate fast), and because most of their value is accrued in the intelligence and application layers, not in the "where does the datacenter sit" layer
fair point on mobility. but the analogy isn't about the physical asset, it's about the capital barrier. a $10B data center is mobile in theory and immovable in practice when only a handful of entities can write that check. the gulf states aren't buying geography, but position in the capital stack.
you're right that value accrues in the intelligence layer today. but that's what refiners said about crude too. the substrate becomes leverage when demand outpaces supply, and every frontier lab is currently power-constrained.
There's no "capital barrier". People/institutions will invest if they expect good returns. Compute just isn't oil, despite people saying it is. The structure of the 2 are just completely different.
the barrier is physics, not capital. you can raise $10B for a data center tomorrow. you can't write a check to thermodynamics. power grids, chip fabs, permitting. none of them care about your IRR. every frontier lab right now is power-constrained, not capital-constrained. that's the oil parallel: the scarce resource was never money, but the ability to physically extract.
That makes the constraint energy, not compute. And the price of solar energy is dropping like a rock as we speak.
BTW, permitting isn't physics. Chip fabs isn't physics (simply money + talent + time; it's not like we will run out of silicon, well, ever). And nothing says data centers can't become more efficient (in fact, they have been) so even power grids and energy aren't all that constraining. There's just nothing in the new solar-fueled AI world that has characteristics like oil.
you just made my point. if the constraint is energy, not compute, then whoever has the cheapest energy owns the bottleneck. that's the gulf states.
solar is dropping in price per panel. deploying it at grid scale for 24/7 baseload is a different problem. data centers don't run on intermittent power. storage, transmission, reliability. that build-out is measured in decades, not quarters.
"chip fabs aren't physics, just money + talent + time." that's also true of oil drilling. the point isn't that the barriers are literally thermodynamic. but that they're hard enough to create durable advantage even when capital is abundant. TSMC spent $40B and 3 years on one fab. the talent pool is measured in thousands globally. "just money + talent + time" is doing a lot of heavy lifting there.
the oil parallel isn't permanent. it doesn't need to be. 10-15 years of infrastructure advantage during the transition is where the rent gets extracted. gulf states don't need compute to be oil forever. they need it to be oil long enough.
Tomas, you forgot to mention India! They'd probably benefit even more than Europe from a switch from fossil fuels to solar energy. Actually, Africa too. In the future, people will look back and call the 21st century the century of China and India (and possibly Africa).
It's a problem for others if their violence boils over and affects others, or triggers a mass of refugees heading towards Europe. After all, we've already seen that happen!
As an environmental consultant dealing with wetlands and forestry for many decades, I am worried about the negative environmental effects of wind power, especially from offshore windfarms.
Not all green technologies are benign, and the devestating impacts on wildlife from wind power are not insignificant.
can't they just use oil for themselves ? Like, building stuff they need and power them with oil instead of selling their oil and getting stuff in exchange ?
Yes, that is a fair point. If their oil is expensive to extract, however, they wouldn't be competitive globally in many areas in many ways. But it does give those countries with cheap-to-extract oil an internal energy resource.
I mean, it still would allow you to do stuff. Also, the Middle East is a great place for harvesting solar energy. A bunch of countries there won't be able to gain a lot of foreign reserves from sales of fossil fuels in a couple decades, but abundant energy might still cushion them.
Yeah but there are plenty of other places that are as good as the Middle East for solar, and if they're not, they're just like 50% worse. Not infinitely worse (as in, the ME has oil and other places don't)
This is 2035 — not 2050. We are probably already at peak oil and the decline of fossil fuels will decrease nonlinearly due to compounding improvements in solar and batteries. The extent that some countries can escape this through diversification (UAE) is a relevant and interesting question.
Worth pointing out that in many cases, a country’s lack of rare earth mineral refining comes from deliberate environmental commitments. The US did used to have its own mines and refineries; we shut them down due to pollution and climate commitments. But the minerals are still in the ground. You can bet as lack of REEs becomes more and more of a national strategic threat, Western governments will adjust their standards.
Excellent article - @Tomas Pueyo is among the limited (but growing) number of commentators who really get that the end of the fossil fuel era is now economically inevitable, and this is a valuable geopolitical projection extrapolating from that.
It's becoming increasingly obvious how dependance on oil fuels conflict.
With respect to rare minerals it is the processing that China has taken control of in part by doing it at a loss without regard for their own environment so almost no one can compete. Countries like Canda that have the minerals should be investing in finding ways to process it without the environmental disaster it is right now.
This will now be done for supply chain independence.
I hope so but we can be very slow at building these sorts of things.
This is not a 10-year thing. More like a 2-5y thing I believe
From idea to functioning plant(s) and mine(s) is definitely not a 2 year thing, 5 years maybe but that's if we drop everything and get onto that. China can do the 2 years but they can do things at a loss without concern for the environment or what people want or don't want built near them.
This is a good explainer on where they are at and the challenges.
https://oilprice.com/Energy/Energy-General/No-Missiles-No-Drones-What-Happens-When-Rare-Earths-Stop-Flowing.html?fbclid=IwY2xjawQX5yBleHRuA2FlbQIxMQBzcnRjBmFwcF9pZBAyMjIwMzkxNzg4MjAwODkyAAEerZHEZIWEGUgjwI09fD4INvQmUselxKX0NMFP63KSkp2J0S7r38vHHsAxJiI_aem_12vQKlZZZzwkAsVmRF6zHg
I really value your opinion, as I've mentioned multiple times, but, in this case, I think you're the one doing a lot of wishful thinking, especially in your 2050 geopolitical predictions. And I'm really failing to understand your prediction for Angola. First of all, ethnic conflict in Angola is practically non-existent since the end of the civil war. Also, according to v-dem and eiu, it's currently an hybrid regime, not an authoritarian state (like your graph says). Second, its internal skirmishes are miniscule when compared to Nigeria and Chad (both in number of casualties and the fact that they don't have the Islamic State on their doorstep). Also, the conflict is exclusive to the exclave of Cabinda, which happens to have all the oil reserves, and that's the big reason there's a fight. If the value of oil is gone, Cabinda has even less reason for fighting.
Angola has a lot of problems and, if the value of oil really falls like you say, Luanda might lose its status as one of the richest and most developped cities in sub-saharan Africa, but I don't see the country plummeting into a Civil War (again); and also don't see them becoming a "strong democracy", out of nowhere. If things go well, they'll simply continued slowly democratizing, like they've had in the last 20 years.
Thank you Pedro.
I am indeed quite hopeful, and my optimism biases my hopes. Not sure what forecast is wishful thinking though? The intro is not a prediction, it’s a potential scenario. And I actually don’t wish that! Too much war in there.
On Angola, it’s a country I haven’t dived deeply into yet so I might get it wrong. But its exposure to oil is objectively high, given the cost of its extraction and the share of gov income it represents. Do you really think the country would remain unscathed with a drop in oil demand?
More importantly, I am not arguing that Angola is very exposed mainly because of its existing internal strife. But by the oil exposure.
I can't imagine Iran losing territory to Azerbaijan. I can't see Turkey allowing the emergence of an independent Kurdistan, and much less one where the Iranian Kurds gained influence over the other kurds of the region. The Iraqi Kurds are the only ones with any sort of autonomy and they seem really keen on peace.
Also, as someone who's got to meet a few immigrants in Saudi Arabia, I think you overestimate the fragility of its governance, so I also don't see Hejaz seceding from it (even with the arguments given in the very insightful article on it).
Syria being at peace with Israel seems extremely hard to achieve (much less in a defense pact under them), because Israel is never gonna trust them enough to give them the Golans back and they simply won't accept that. I don't know how their story end, but I'm pretty sure it's not with Syria bowing to Israel.
And why would Dubai, with the strongest passport in the world, one of (if not the) most important airport, and one of the highest soft power capabilies in the world, throw away its stability and risk giving it all up for a petty war? Even if they had anything to gain by it, the losses they would get from the war would make them reconsider.
As for Angola, a drop in oil demand will definitely affect the country, but I think only Luanda and Cabinda would notice. And I simply don't see how that could ever lead to a civil war. Who would be fighting whom? For control of what?
Yes, of course none of that is a prediction! This is one of millions of potential scenarios! The idea is to paint a picture of the mess that could happen with the end of oil, to make it real. And I sure believe that's the case: The transition will be messy!
Syria is at peace with Israel at this very moment.
I would argue they are simply taking care of other affairs first, but the Israel situation is not forgotten.
How many people predicted a (bloody, long) civil war in Libya in 2010? Heck, how many people predicted a (bloody, long) civil war in Syria in 2010? Or Yemen?
The break up of a bunch of countries in the Middle East actually isn't as unlikely as you seem to think (granted, the various emirates of the UAE going to war with each other seems a bit farfetched).
Where’s the role of nuclear power in this analysis?
I’m a mega supporter of nuclear, but it takes a long time to be built. Regardless, whether it’s through solar or nuclear, the reaction to oil doesn’t change.
A bit more conventional, but with great sources:
https://am.jpmorgan.com/us/en/asset-management/liq/insights/market-insights/eye-on-the-market/energy-paper-2026
There are several tax haven countries that provide the equivalent of pumping regulatory arbitrage out of the ground. The smaller petro states could become this, but the emphasis is on small. You can't share the resource with tens of millions of people.
Hence the UAE
Really interesting article
However, you only touched on oil as being an energy source.
But isn't plastic and other industries also reliant on oil? Do you expect those to not matter in the future?
I think plastics are something like 6% of oil and none of gas? If so, dividing by 20x the amount of oil sales in the world is close enough to eliminating it in geopolitics terms.
I'd be delighted to hear how the Iran conflict impacts on how this article shakes out. in both "worlds" China is the big winner .
it’s been awhile since I heard the words ‘Peak Oil’ on the internet.
I guess we can slot this in now that ‘Climate Change’ is no longer an existential threat?
I think 'peak demand' is more appropriate.
Yes, peak not as in supply peak but demand peak
Climate change is no longer an existential threat exactly because battery technology is causing peak oil
the diversification framing undersells what's actually happening. when MGX targets $100bn in AI infrastructure and Saudi PIF combines $240bn for data centers, the story isn't petrostates escaping rent-seeking, but petrostates porting the rent model to a new layer.
oil worked because it was physically concentrated, capital-intensive, and infrastructure everybody had to use. compute is shaping up identically. the gulf states know the playbook because they built the last version of it.
norway escaped the resource curse by investing the rent into diversified global assets. most of the others are shopping for the next extractable moat. i wrote about the AI landlord structure (https://huliwood.substack.com/p/the-ai-landlords). your 2050 scenario adds the layer i'd underweighted: the most dangerous outcome isn't petrostates failing to diversify. rather: petrostates discovering that compute IS the new oil, and they already have the capital to buy the wells.
There is no way AI can be exploited in the same way as oil, because datacenters can move (and depreciate fast), and because most of their value is accrued in the intelligence and application layers, not in the "where does the datacenter sit" layer
fair point on mobility. but the analogy isn't about the physical asset, it's about the capital barrier. a $10B data center is mobile in theory and immovable in practice when only a handful of entities can write that check. the gulf states aren't buying geography, but position in the capital stack.
you're right that value accrues in the intelligence layer today. but that's what refiners said about crude too. the substrate becomes leverage when demand outpaces supply, and every frontier lab is currently power-constrained.
There's no "capital barrier". People/institutions will invest if they expect good returns. Compute just isn't oil, despite people saying it is. The structure of the 2 are just completely different.
the barrier is physics, not capital. you can raise $10B for a data center tomorrow. you can't write a check to thermodynamics. power grids, chip fabs, permitting. none of them care about your IRR. every frontier lab right now is power-constrained, not capital-constrained. that's the oil parallel: the scarce resource was never money, but the ability to physically extract.
That makes the constraint energy, not compute. And the price of solar energy is dropping like a rock as we speak.
BTW, permitting isn't physics. Chip fabs isn't physics (simply money + talent + time; it's not like we will run out of silicon, well, ever). And nothing says data centers can't become more efficient (in fact, they have been) so even power grids and energy aren't all that constraining. There's just nothing in the new solar-fueled AI world that has characteristics like oil.
you just made my point. if the constraint is energy, not compute, then whoever has the cheapest energy owns the bottleneck. that's the gulf states.
solar is dropping in price per panel. deploying it at grid scale for 24/7 baseload is a different problem. data centers don't run on intermittent power. storage, transmission, reliability. that build-out is measured in decades, not quarters.
"chip fabs aren't physics, just money + talent + time." that's also true of oil drilling. the point isn't that the barriers are literally thermodynamic. but that they're hard enough to create durable advantage even when capital is abundant. TSMC spent $40B and 3 years on one fab. the talent pool is measured in thousands globally. "just money + talent + time" is doing a lot of heavy lifting there.
the oil parallel isn't permanent. it doesn't need to be. 10-15 years of infrastructure advantage during the transition is where the rent gets extracted. gulf states don't need compute to be oil forever. they need it to be oil long enough.
Tomas, you forgot to mention India! They'd probably benefit even more than Europe from a switch from fossil fuels to solar energy. Actually, Africa too. In the future, people will look back and call the 21st century the century of China and India (and possibly Africa).
That’s tTue!
The Saudis eating sand and Muslims killing each other with rocks. Why try to prevent this happy state of affairs?
It's a problem for others if their violence boils over and affects others, or triggers a mass of refugees heading towards Europe. After all, we've already seen that happen!
Put in another way: “even if you think you want that, you don’t want that. Look what happened with Syria.”
As an environmental consultant dealing with wetlands and forestry for many decades, I am worried about the negative environmental effects of wind power, especially from offshore windfarms.
Not all green technologies are benign, and the devestating impacts on wildlife from wind power are not insignificant.
Thankfully, solar is likely to be much more viable than wind!
Say you don’t believe in human-caused climate change without saying you don’t believe in human-caused climate change.
Great article but I have question :
can't they just use oil for themselves ? Like, building stuff they need and power them with oil instead of selling their oil and getting stuff in exchange ?
Yes, that is a fair point. If their oil is expensive to extract, however, they wouldn't be competitive globally in many areas in many ways. But it does give those countries with cheap-to-extract oil an internal energy resource.
Oil is useful to do stuff.
If you can't do stuff as well as others, even if your energy is available and cheap, it's meaningless
I mean, it still would allow you to do stuff. Also, the Middle East is a great place for harvesting solar energy. A bunch of countries there won't be able to gain a lot of foreign reserves from sales of fossil fuels in a couple decades, but abundant energy might still cushion them.
Yeah but there are plenty of other places that are as good as the Middle East for solar, and if they're not, they're just like 50% worse. Not infinitely worse (as in, the ME has oil and other places don't)